【作者】 张芳；

【导师】 张小曳； 周凌晞；

【作者基本信息】 中国气象科学研究院 ， 气象学， 2011， 博士

【摘要】 论文主要对瓦里关多年大气CO₂、CH₄和CO浓度资料进行了再分析及源汇研究,所使用的CO₂、CH₄和CO浓度资料的时间段分别为1995²008年,2002²006年和2004年7月²007年6月。根据历史标定比对实验结果的再分析及再标定实验,校正和更新了CO标准尺度;对多级质量控制方法对有效数据筛选的影响进行了探讨,研究建立了数据处理和质量控制方法,并将大气CO₂、CH₄和CO浓度时间序列分别更新至X2007、NOAA04和WMO2000的最新尺度。利用局部近似回归法对大气CO₂、CH₄和CO浓度进行本底值筛分,获得1995²008年间CO₂浓度的本底数据、源抬升数据和吸收汇数据（受陆地生态系统植被的吸收的影响）分别占了72%±5%、17%±4%和11%±2%;从1995²000年到2004²008年,源抬升浓度百分比由14¹5%上升至18%¹9%,说明瓦里关大气CO₂浓度受到人类活动的影响正在逐年增强;吸收汇的数据量比例变化相对稳定（10¹4%）,反映了近几年瓦里关地区陆地生态系统对CO₂吸收汇的作用相对稳定。大气CH₄和CO浓度本底数据百分比分别占⁵8%和⁴3%,表明在观测期间受到较强的区域源和汇的影响。1997²008年的5d后向轨迹（500hPa）分析表明,影响瓦里关的气流主要来自西部地区,但是夏季主要受到来自东部或东南部气流的影响,且因该类轨迹途径人口密集地区,>90%的为污染气团。典型个例分析显示,不同季节大气CO₂、CH₄和CO浓度高值与来自青海西北部（尤其是格尔木地区）和瓦里关东南或东部（西宁和兰州一带）的空气团轨迹关系密切;其浓度低值则对应于来自西藏西北部、青海和新疆南部地区的空气团。轨迹聚类统计分析结果也呈类似特征,2002²007年大气CO和CH₄浓度高值受来自瓦里关东南或东部地区（西宁、兰州一带）气团的影响（夏季尤其明显）,可看作污染扇区;而春、秋和冬季则以来自西北方向的气流为主导,约50%该来向的气团轨迹会抬升CO和CH₄浓度,因此并不单纯为污染或清洁气流;来自西藏和青海省南部的气团轨迹对CO₂、CH₄和CO浓度的抬升几乎无贡献,可看作清洁扇区。本研究中值得注意的是几乎所有来自格尔木地区的轨迹都会使大气CO₂、CH₄和CO浓度抬升,说明中国西部地区经济发展和能源结构改变对温室气体浓度时空分布的影响;另注意到夏季来自甘肃或宁夏黄河沿岸的农业区（如水稻种植区）气团会使CH₄浓度抬升明显,显示了CH₄的农业种植来源。冬季瓦里关大气CO₂、CH₄和CO的抬升浓度两两之间呈显著的正相关（r2>0.7, p<0.01）,说明3者具有相同的来源（如化石燃料和生物质燃烧）。尝试利用CO比值相关法估算的2005年和2006年中国CO₂排放量分别为6214.7 Mt和6320.6Mt,与清单统计方法估计的结果接近。CH₄排放量分别为16.44 Mt （2005年）和20.11 Mt（2006年）。大气CO₂、CH₄和CO多年日变化其季节差异性方面,1995²008年冬春季节大气CO₂浓度呈白天高（高值出现在11:00¹5:00）,早晨和夜间低的特征,主要是由于植物的光合作用减弱,白天周边居民活动及土壤呼吸释放CO₂等使白天浓度高于夜间;研究期间夏季和秋季由于白天陆地生态系统植物吸收CO₂的汇增强,且对流旺盛,扩散条件好使其浓度有所稀释,16:00¹8:00出现一天之中的最小值。CO₂浓度的平均日振幅夏季最大（2.3ppm）,春季和冬季最小（都为⁰.7ppm）,主要与陆地生态系统在不同生长周期与大气之间气体交换的差异有关。大气CH₄（2002 ²006）和CO（2004年7月²007年6月）浓度多年日变化规律与CO₂类似,但是夏季CH₄和CO浓度在下午较低的原因与CO₂有所不同,前2者主要由于它们与OH的光化学作用很强而导致其浓度较低,而后者则是由于植物的光合作用导致。季节变化方面,1995²008年瓦里关大气CO₂浓度夏秋低（8月最低）,冬春高（4月最高）,与美国夏威夷Mauna Loa（155.57W , 19.52N , 3397m asl）、德国Schneefernerhaus（47.25N, 10.59E, 2962 m asl）和美国Niwot Ridge（40.05N,105.59W, 3523m asl）总体变化趋势类似,但有一定的相位差异,主要受各站下垫面特征差异及其所在区域源汇相互作用的强弱的影响;2002²006年,瓦里关大气CH₄浓度在夏季出现高值（6⁸月达到最大值）,冬春较低,而Mauna Loa和Niwot Ridge则完全相反。这是夏季瓦里关地区排放源增强（居民放牧增多）和来自东南部地区污染气团输送的双重作用的结果;而Mauna Loa和Niwot Ridge纬度较低,夏季CH₄的光化学作用占主导而使其浓度较低。2004年6月²007年7月,CO浓度春季（3月和4月）最高,秋季（9月和10月）最低,与瑞士Jungfraujoch（46.55N, 7.99E, 3580m asl）、Niwot Ridge和Mauna Loa大体趋势基本一致。瓦里关大气CO₂、CH₄和CO多年平均季振幅分别约为9.0ppm、11ppb和26ppb,与Mauna Loa、Niwot Ridge和Schneefernerhaus的季振幅有所不同。变化趋势方面,大气CO₂本底浓度呈逐年快速增长,瓦里关大气CO₂本底浓度增长率从1991¹995年的+1.3 ppm yr-1（1991¹994年的CO₂数据来自瓦里关flask瓶采样数据）增长到2000²004年的约+2.4ppm yr-1,与全球平均状况类似。分别采用以干洁自然大气配制和工业来源的标气进行测试（其CO₂浓度范围为350⁷00ppm）,探讨不同技术测量大气CO₂浓度时的同位素效应,初步结果表明,当del13C值在自然丰度范围时,其测量偏差△CO₂绝对值最小（|△CO₂|<1.0ppm）;测量工业源的标气样品（del13C为-9^-28‰）时其偏差较大（可高达10ppm）。发现当标气样品的CO₂浓度较高时,系统之间的测量偏差呈非线性增加,推测CO₂浓度太高和del13C值太负均有可能导致系统之间测量偏差较大。不同技术测量CO₂的同位素效应的定量校正尚需进一步的研究。更多还原

【Abstract】 Mixing ratios of atmospheric carbon dioxide （CO₂）, methane （CH₄） and carbon monoxide （CO） observed at Mount Waliguan （WLG）, a global background station in remote western China, were re-analyzed and evaluated. The data periods are 1995²008, 2002²006 and July 2004^June 2007 for CO₂, CH₄ and CO respectively. Corrections for drift in reference gases were also included in the data revision according to the history calibrations and intercomparison experiments. The measurement data of CO₂, CH₄ and CO at WLG have been updated to latest X2007, NOAA04 and WMO2000 scale respectively. Data treatment and quality control procedure has also been established to improve scientific use and application of the data. A mathematical procedure based on robust local regression was applied to distinguish background and non-background data, as is actually to classify the impact of regional emissions or influence of polluted air parcel. From 1995 to 2008, approximately 72%±5%, 17%±4% and 11%±2% of all observed CO₂ data have been selected as background, polluted and sink data, respectively. The percent for polluted CO₂ data increased from 14-15% in 1995-2000 to 18-19%, reflecting enhanced impact from human activities （e.g. fossil fuel emissions） in recent years; the percent of data representing CO₂ sinks didn’t change much （10-14%） indicating a relatively constant uptakes from terrestrial ecosystems in the region. However, the more polluted （elevated） than the uptakes CO₂ suggested an imbalance between emissions and sinks for atmospheric CO₂ there during the past 15 years at WLG. By this method, about 58% and 52% of all data were selected as CH₄ and CO background data, respectively, indicating significant influences from regional emissions/sources.By analyzing 5d back-trajectory （500hPa） of 1997²008 at WLG, it showed air parcels arriving at WLG were predominately from the west, except in summer when advection from the east and southeast prevailed. >90% of trajectories from the east typically brought polluted air to the site, having passed over populated urban areas upwind. The case study combined with trajectory analysis showed the episode of high CO₂, CH₄ and CO mixing ratios were associated with advection from the heavily populated regions east or southeast （e.g. Xining and Lanzhou） of WLG and northwest of Qinghai via Ge’ermu urban area where growing industrial activities as well as crops residue burning provide large sources of CO suggesting a large source area due to human activities; whereas, the low values were observed most frequently when air masses originated from the sparsely populated Tibet and south of Qinghai and Xinjiang Uygur Autonomous Region （XUAR）. By combining the observed data with trajectory-based statistics and cluster analysis, it has been demonstrated that air masses originating from the east to southeast of WLG have the strongest impact on CO and CH₄, however, only in summer transport from this direction significant. In the other seasons, air parcels arriving via the northwest are more common. These exhibit both background and polluted characteristics. Air from the central XUAR and the Ge’ermu urban area have show enhanced CO and CO₂ levels due to the growing economy in west China as well as biomass burning in the region. Air parcels coming from the sparsely populated Tibet contribute least to enhance all CO₂, CH₄ and CO values observed at WLG. It should be noticed that mixing ratios of atmospheric CH₄ would be enhanced when air masses originated from Ningxia and northwest of Gansu agriculture areas （mainly growing rice） along the Yellow river region, especially in summer. The probability that air parcels pass over regions of clean or polluted regions were further identified using potential source contribution function （PSCF） analysis, and all the CO₂, CH₄ and CO displayed similar source region distribution as above described. Intercorrelation of atmospheric CO₂、CH₄ and CO above the background showed a very well correlation （r2>0.7, p<0.01） in winter probably in account of their common sources being mainly from fossil fuel and biomass burning. Hence, CO₂ and CH₄ emissions from fossil fuel have been estimated by using CO ratio method. Emissions for CO₂ in China were 6214.7 Mt in 2005 and 6320.6Mt in 2006 respectively, which is similar to the results by statistical method. Emissions for CH₄ in China were 16.44 Mt in 2005 and 20.11 Mt in 2006.Regarding diurnal variation of atmospheric CO₂ mixing ratios based on observed data from 1995 to 2008, it has been found that atmospheric CO₂ showed higher during daytime and lower in early morning and nighttime in spring and winter, the highest mixing ratios most frequently occurred at 11:00¹5:00 when residents’activities （i.g. grazing, heating） increased, but in the cold season photosynthesis is very weak and make no significant impact on CO₂ levels. Whereas mixing ratios of CO₂ exhibited daytime minimum and nighttime maximum in summer and autumn, the lowest CO₂ mixing ratios presented at 16:00¹8:00, as mainly due to the enhanced uptake of CO₂ by vegetations in summer in north hemisphere. In addition, the meteorological conditions of strong convection in summer are also advantageous to pollution diffusion and will decrease CO₂ mixing ratios during daytime. Meanwhile, effects of temperature inversion during evening or early morning at high altitude site like WLG also would accumulate CO₂ mixing ratios. Averaged diurnal variation of CO₂ displayed largest in summer with amplitude of 2.3ppm （part per million） but only with ⁰.7ppm both in spring and summer, suggesting differences in carbon exchange between terrestrial ecosystems and atmosphere attribute to plant growing cycle over one year. Diurnal variation of atmospheric CH₄ and CO based on data during 2002²006 and during July 2004^June 2007 respectively showed a similar pattern with CO₂ in all seasons, but with different causes of the lower mixing ratios for CO and CH₄ during daytime in summer. As we known, it’s because of photosynthesis for lower CO₂ mixing ratios in the afternoon; however, the mainly impelling force is photochemical reaction with OH radicals which would decrease mixing ratios of CH₄ and CO during daytime in summer.The averaged CO₂ seasonal cycle from 1995 to 2008 at WLG displayed April maximum and August minimum, declining rapidly in late spring and early summer （from May to June） and increasing in autumn （from September to November） mainly caused by plant and soil respiration as well as plant photosynthesis. The overall characteristic is similar to that of observed at Mauna Loa, Schneefernerhaus and Niwot Ridge, but exhibiting certain phase delay mainly influenced by the differences in interaction of regional sources/sinks as well as the underlying surfaces. The averaged CH₄ seasonal cycle during 2002²006 showed minimum in spring and winter, maximum in summer （from June to August）, as is totally opposite to that of observed at Mauna Loa and Niwot Ridge. The high levels of CH₄ at WLG should be attributed to enhancing regional/local sources （i.g. herd） around the site as well as the dominant polluted air flow from southeast region in summer. Another possible reason is, compared to Mauna Loa and Niwot Ridge, the photochemical effect at WLG （located at higher latitude） is weaker, as means CH₄ sink is smaller than that of at these sites with lower latitude in summer. Seasonal cycle of atmospheric CO mixing ratios from July 2004 to June 2007 exhibited maximum in spring （March and April） and minimum in autumn （September and October）. This is in good agreement with the variability observed at Jungfraujoch, Niwot Ridge and Mauna Loa. Averaged seasonal amplitude of CO₂, CH₄ and CO during the observed periods was 9.0ppm, 11ppb （part per billion） and 26ppb, respectively. Mixing ratios of atmospheric CO₂ increased exponentially, as is well agree with the trend observed at Mauna Loa. The growth rate of has accelerated since measurements began at WLG in 1991 where CO₂ increased from nearly 1.3 part per million per year （ppm yr-1） in 1991-1995 （the data before 1994 obtained from flask measurement at WLG） to 2.4ppm yr-1 in the recent years, but with large year to year variations, as is closely associated with the climate events （e.g. El Nino）.Preliminary results of isotope effect on measuring CO₂ mixing ratios by using different techniques or systems showed that the differences are within±1.0ppm for the samples with natural abundances del13C（-7.7^-9.0‰） of CO₂; The differences could increase to 10ppm when del13C decrease to about -25.0‰. And also we found that both high CO₂ concentration and low del13C can lead to large measurement bias by different techniques. However it is still need further studies to demonstrate this conclusion and to calibrate the bias due to the istope effect qutitativly.更多还原